Human stories without humans: Has AI degraded China’s micro-dramas?

(Edited and refined by Candice Chan, with the assistance of AI translation.)

After 15 straight days on the job, Qing (a pseudonym) was counting down to his first day off. The video editor at an AI micro-drama company in Hangzhou had spent the previous 14 nights working until at least 11pm, as his three-person team struggled to keep up with a weekly quota of 70 episodes.

“To put it bluntly, we’re burning our lives away,” he told Lianhe Zaobao. “I’m planning to quit.”

AI-generated micro-dramas have emerged as one of the hottest trends in China’s film and television industry in recent months. Using increasingly accessible AI tools, creators can now produce complete episodes with visuals, dialogue and sound effects in a fraction of the time and with far fewer people than traditional productions require.

The breakout success of AI-generated historical dramas such as Huo Qubing (《霍去病》), based on the life of the Western Han dynasty general, has fuelled expectations that the technology could unlock a vast new commercial market.

However, the costs of the AI micro-drama boom have quickly become apparent. The live-action micro-drama industry, which was still expanding as recently as last year, has suddenly found itself losing market share, leaving actors facing unemployment. Meanwhile, the unchecked growth of the AI micro-drama sector has begun to create problems of its own.

On the demand side, AI micro-dramas are facing growing scrutiny over their quality and originality. Behind the scenes, the industry’s cut-throat competition has fuelled complaints of overwork, burnout and destructive rivalry. Qing’s experience offers a glimpse into the human cost of the AI micro-drama boom.

Quantity over quality

Qing joined an AI micro-drama company in March, where he produces content for multiple distribution platforms. He earns 5,000 RMB (US$735) a month, a salary he says bears little relation to the workload.

He explained that each AI micro-drama has a production cycle of only five to seven days. “A new project starts on Monday, and the finished product has to be delivered by Sunday. Without working overtime, it’s simply impossible.”

The compressed production schedule is driven by a simple business logic: maximise output in the hope of producing a hit.

“The quality requirements for AI micro-dramas aren’t particularly high — the key is speed. The more dramas you produce, the greater the chance that one becomes a blockbuster.”

As large numbers of fortune-seekers have poured into the AI micro-drama market, concerns over the overcapacity typical of emerging industries have begun to surface.

According to Chinese media reports, more than 2,100 AI micro-drama companies have been registered in China this year alone, yet genuine hits remain exceedingly rare. Out of over 120,000 AI micro-dramas released by February this year, fewer than 150 surpassed 100 million views, representing a blockbuster rate of roughly one in a thousand.

As a result, companies have resorted to every possible means to stay competitive, intensifying the industry’s race to the bottom. In March, after the release of the Seedance 2.0 video generation model, an AI micro-drama company in Changsha extended employees’ working hours to 3am so they could secure off-peak computing resources, sparking widespread public criticism.

The real winners

These punishing working hours stand in stark contrast to the industry’s meagre profits. Chen Jian, chairman of Xihongshi Films (西红柿影业), which entered the AI micro-drama market last year, admitted when interviewed: “For companies producing AI micro-dramas on contract, there’s very little profit left after paying for computing power and staff costs. The money mostly goes to the leading tech giants. We’re more like sweatshops.”

Chen noted that production fees have continued to fall, dropping from around 4,000 RMB per minute in the early days to just 800 RMB per minute today.

The immaturity of AI technology is another reason margins remain so thin. Chen said: “Sometimes you issue 100 prompts, and only one produces the result you’re looking for... We thought we’d make money, but it turned out we were losing it because endless revisions drove up both labour and computing costs.”

While AI has undoubtedly improved productivity, it has not fundamentally altered the industry’s entrenched logic of fierce competition and unequal profit distribution. Many practitioners have found that the real winners are the platform companies and a handful of leading producers, while the vast majority of small and medium-sized production houses struggle merely to survive.

An assembly-line approach to creativity

For others in the industry, the “sweatshop” label also reflects an increasingly assembly-line style of production.

Unlike the creative development process behind traditional film and television productions, many AI micro-dramas are produced through a highly standardised workflow. From script selection and image generation to post-production, every stage is broken down into measurable tasks in pursuit of maximum efficiency.

Having previously worked on live-action micro-dramas, Qing said: “There’s still some creativity involved in live-action productions. AI micro-dramas feel more like assembling products in a factory. If one shot doesn’t work, you simply find another clip to patch it together. There’s not much technical skill involved.”

This mechanised production model is perhaps felt most keenly by so-called “card drawers”, a new occupation that has emerged alongside the AI micro-drama industry. Their job is to repeatedly input prompts into AI models to generate plenty of material before selecting usable images. Because AI output is inherently random, they must constantly refine their prompts until they successfully “draw” a scene that meets the requirements.

Huang Tao (pseudonym), who works as a card drawer at an AI micro-drama company in Jiangxi province, said the process can be maddening.

One of the central debates surrounding AI micro-dramas is whether this increasingly industrialised production model will further homogenise content and ultimately reduce what has traditionally been a creativity-driven film and television industry into a manufacturing business.

Factory production means maturation?

Huang believes whether factory-style production is appropriate ultimately depends on audience expectations.

“If viewers don’t care too much about quality, then factory production is feasible. But if you want to create genuinely outstanding productions, then it definitely won’t work.”

Xu Jian, a professor at Shanghai Jiao Tong University’s School of Media and Communication, takes a different view. He believes more specialised divisions of labour are not inherently problematic; what matters is whether the story itself is compelling.

He told LHZB that greater specialisation improves consistency, reliability and predictability in the final product. In fact, what people describe as “factory production” reflects the maturation of the AI micro-drama industry.

“It no longer depends on individuals. There is capital investment, revenue generation, and people can make money from it.”

Signs of contraction

In just six months, the AI micro-drama industry has shifted from explosive expansion to a period of consolidation. The sweatshop model — characterised by wafer-thin profits, high-pressure working conditions and talent attrition — has led many practitioners to realise that simply increasing production volume is not a sustainable strategy.

The market has already begun showing signs of contraction. According to big data platform DataEye, around 39,500 AI micro-dramas or AI comic dramas were uploaded to Douyin in May this year, roughly 10% less than in April.

Xiaoyu (pseudonym), a scriptwriter at an AI micro-drama company in Hangzhou, believes many poorly made productions are now being eliminated by the market.

She said many newcomers with no industry background have recently entered the business, inevitably affecting quality. “Aesthetic standards have become distorted.”

She complained: “Many AI micro-dramas go for flashy special effects and visual spectacle, it’s basically mindless entertainment. Some male-oriented dramas can’t even be bothered writing a proper comeback story anymore — the protagonist is already overpowered in the first episode and stays that way until the last. The plots are completely devoid of substance.”

Xiaoyu believes that as the industry becomes more regulated, many companies that entered the market simply to chase the trend will eventually disappear.

“I’m prepared to be laid off at any time,” she admitted.

More than a passing fad

Facing increasingly challenging market conditions, production company owner Chen Jian is searching for a new path forward. He acknowledged that although his company is currently profitable, those earnings are still “sweatshop money”, and profits remain about 30% lower than when the company was producing traditional film and television content.

Chen revealed that the company is gradually shifting away from producing dramas solely on commission for clients and towards creating and distributing its own productions. If successful, the profits would flow directly back to the company.

Despite the controversies surrounding AI micro-dramas, both Chen and Professor Xu believe the genre is far from being a passing fad. Whenever new technologies disrupt established industries, scepticism and resistance inevitably follow, and AI micro-dramas are no exception.

Xu believes AI micro-dramas will eventually move towards higher-quality productions. As AI technology becomes more widely accessible, the decisive factor will once again be the quality of storytelling.

He cautioned: “Everyone entering this industry should avoid becoming dazzled by AI technology or getting carried away by how impressive it seems. Don’t fall into the trap. Ultimately, the core competitive advantage of AI micro-dramas will still come down to the ability to tell a good story.”

He drew the analogy: “The race has already begun. The thoroughbreds and the dark horses will only emerge through competition and fierce rivalry… We should have confidence in the market’s ability to cleanse itself.”

This article was first published in Lianhe Zaobao as “AI短剧狂飙催生“血汗工厂” 流水线上的内卷与突围赛”.

Related: [Big read] Can Shanghai turn ‘vulgar’ micro-dramas into premium entertainment? | ‘Brainless’ but satisfying: How Chinese micro-dramas are taking over our screens